Dataset Torrent

Dataset Card for Audius

Dataset containing Audius archives.

• audio-classification
• audio
• metadata
• multilingual

Dataset Summary

This dataset contains metadata for 716,340 audio tracks from the Audius decentralized music platform. Audius is a blockchain-based streaming service that allows artists to share and monetize their music directly with listeners. The dataset includes detailed metadata about tracks, artists, and engagement metrics.

Languages

The dataset is multilingual with English as the primary language.

Dataset Structure

Data Fields

The metadata for each track includes artwork URLs, description, genre, unique identifiers, content identifiers, original filename, availability flags, mood, release date, remix information, repost count, favorite count, comment count, tags, title, user information, duration, download permission, play count, permalink, stream availability, playlist containment, and pinned comment references.

Data Splits

All tracks are in a single split.